Skip to content

stepanogil/assistants-api-hr-copilot

Repository files navigation

HR AI Copilot built using the Assistants API

Description

This is an an AI-powered HR Assistant powered by GPT-4-Turbo. It integrates a suite of tools including SAP HR for employee data retrieval, ServiceNow for ticketing information, GPT4 Vision for OCR capabilities and Code Interpreter for data analysis.

It uses Streamlit as frontend.

Tech Stack

  1. Assistants API: Orchestrates the overall system, managing interactions and data flow.
  2. GPT-4-Turbo: Acts as the 'brain' of the system, routing to appropriate tools, synthesizing user input and tool output, and generating responses.
  3. SAP HR (via API): Serves as a data source for answering user queries about their own employee data.
  4. Azure SQL: Provides data for the time data of user's direct reports. Integrates with SAP HR for data pipeline.
  5. Code Interpreter: Facilitates data analysis and visualization (e.g., using Python libraries like pandas, matplotlib) on the time data.
  6. Azure Cognitive Search (Azure AI Search): Functions as the vector database for HR Policy documents.
  7. LangChain: Utilized for the vector store retriever within Azure AI Search.
  8. GPT4 Vision: Employs Optical Character Recognition (OCR) for processing receipts.
  9. ServiceNow (via API): Acts as a data source for updates on ServiceNow tickets.
  10. Streamlit: Provides a reliable and lightweight user interface.

Installation

# Clone the repository
git clone <repository-url>

# Navigate to the project directory
cd <project-directory>

# Install required Python packages
pip install -r requirements.txt

# Set up environment variables
# Create a .env file and populate it with necessary API keys and IDs

# Run the Streamlit app
streamlit run app.py

Usage

  1. Start the Streamlit application.
  2. Click 'New Thread' to initiate a conversation with the AI HR Assistant.
  3. Type your HR-related queries or commands and receive responses or actions from the assistant.
  4. The assistant can process images for OCR, fetch employee data, and handle ticket information.

Demo - click to watch the video

Watch the Video

Author

Stephen Bonifacio

Feel free to connect with me on:

Linkedin: https://www.linkedin.com/in/stephenbonifacio/
Twitter: https://twitter.com/Stepanogil

Did you find this useful? If you'd like to show some love, I won't say no to a cup of coffee! 🤗

ko-fi

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors